Opera 9.5 Beats Firefox and IE7 As Fastest Browser

Abhinav Peddada writes "Ars Technica takes Opera 9.5, the latest from Opera's stable, for a test run and finds some interesting results, including it being a 'solid improvement to an already very strong browser.' On the performance front, Ars Technica reports 'Opera 9.5 scored slightly higher (281ms) than the previous released version, 9.23 (546ms). And Opera 9.x, let it be known, smacks silly the likes of Firefox and Internet Explorer, which tend to have results in the 900-1500ms range on this test machine (a 1.8 GHz Core 2 Duo with 2GB RAM). Opera was 50 percent faster on average than Firefox, and 100 percent faster than IE7 on Windows Vista, for instance.'"

Wasn't that always the case?

From what I've seen the speed rankings in all tests always have Opera and Safari leading with IE and FF being behind.

Re:Different market

From what I've seen the speed rankings in all tests always have Opera and Safari leading with IE and FF being behind.

Opera aims at different market -- small gadgets. This is where the speed is really critical. For IE and FF good enough is enough, since performance on modern desktops is not that critical.

From what I've seen the speed rankings in all tests always have Opera and Safari leading with IE and FF being behind.

Opera aims at different market -- small gadgets. This is where the speed is really critical. For IE and FF good enough is enough, since performance on modern desktops is not that critical.

As a Quad G5 (4x 2500) Mac owner with lots of RAM, I really don't want a browser choking up an entire CPU and flooding my memory. I didn't pay money to cover amateur programming mistakes by other people. As same guy, I flamed Opera guys about not fixing a bug happens on Slashdot beta, first thing I checked was that after getting that awesome 9.5 alpha and yes it is fixed.

I have used a Xeon Video workstation lately and poor AVID was acting like it is on 80386 because a stupid "free" antivirus was taking whole CPU cycles trying to "scan" gigabyte level raw videos while it was asked to ignore them.

It is common getting replies as "get more RAM" or "upgrade your CPU" from various browser fans but when I see a browser using 100% CPU , I get alerted about what kind of security issues it may have and why I should be wasting my CPU to it.

Opera's power comes from managing to code and sell full feature browsers which would even run on Nokia 7650 with 2 MB of RAM. Don't let the Desktop versions memory usage fool you, it is mostly RAM Cache, not memory "flood". Instead of flooding memory, they use it for a good reason and release immediately when another app needs it.

Re:Different market

I just switched from FF to Opera because of its low market share numbers - which was the same reason I switched from IE to FF when the FF market was about 2%.

Pffft. I'm must more emo than you, I use Lynx which has practucally no market share!

Re:Different market

I really wouldn't say that. Once you've used a browser that renders pages considerably faster than your old browser, there's no going back. It makes a *big* difference.

Yes, it does makes difference, but on desktop feature set is much more important and there is no way I'm trading NoScript + CookieSafe + Firebug + Foxmarks + Slashdotter for a slight increase in speed.

Re:Different market

there is no way I'm trading NoScript + CookieSafe + Firebug + Foxmarks + Slashdotter for a slight increase in speed.

+ Adblock + a few other things, and that 'slight increase in speed' might start to look like a supersonic jet outrunning a kid with a wheelbarrow. A wheelbarrow with a lot of nifty stuff on it, sure, but still ;)

Re:Wasn't that always the case?

The Opera web browser is no longer ad supported. Just thought you should know.

How Opera is Supported

Then how is it supported?

They have deals with search engines, like Google and Yahoo, to get placement as the default engines in the toolbar, in Speed Dial, and in Opera Mini. (I think these days it's Yahoo in all 3.) Same kind of deal that Firefox has with Google, really.

Plus there are the versions for devices [opera.com] (Nintendo DS, etc.), which they still charge for, either directly or through licensing deals with device manufacturers and mobile carriers. So they pull in revenue from that.

This article is a year out of date, but still informative: Opera making big profits from free software [itwire.com.au].

Re:Wasn't that always the case?

You're only 2 years out - the ads were dumped in Opera 8.5, and that was released on the 20th of September 2005.

If you're going to complain about something, please try and make it relevant.

Re:Wasn't that always the case?

Forgive the person for not knowing, he's still waiting for the news to load in one of the slower browsers.

Opera without Pavarotti

Luciano Pavarotti September 6, 2007 R.I.P. yesterday.

http://en.wikipedia.org/wiki/Luciano_Pavarotti [wikipedia.org]

What matters Opera without Pavarotti about?

Re:Wasn't that always the case?

If you're going to complain about something, please try and make it relevant.

A relevant complaint, like having to wait longer for webpages to render?

Maybe I just don't have spiderman senses or Clint Eastwood style reflexes that most web users have, but the wait of less than half a second for a webpage to render doesn't really bother me that much.

I'm not saying this because I'm a Firefox fanboy, or because I don't like Opera, I just don't get why it matters. Even on MySpace it doesn't take so long to render a webpage that it bothers you, and if a webpage takes a long time to load it'll almost certainly be because of your network connection or the server and not rendering time.

Re:Wasn't that always the case?

You may or may not have noticed, but many websites, including this one, are using more and more javascript, which are what the speed tests were measuring (sue me, I RTFA'd). So it's not so much the initial load time, as it is how fast other things happen once the page has been displayed.

I wonder........

I wonder if they would have said this if Pavarotti hadnt just died?

Grade article: incomplete

Well... okay. That was a short article.

I'm not expecting them to try Lynx or anything, but at least test Safari on Windows? The one that also claims to be fast?

Re:Grade article: incomplete

Opera used to have problems with Digg back in the 8.x days, but since 9.x it works just fine.

First post thanks to OPERA!!!!

Those milliseconds really add up...

Re:First post thanks to OPERA!!!!

so i guess it really isnt that fast then?

Article is very misleading - JS benchmark only

The article links to a Javascript benchmark only. There are many many more variables involved in determining how fast a given browser is, although certainly Javascript plays it's part. Variables like how soon does the browser start processing incoming, but yet incomplete data, etc. influence the browser's snappiness a lot aswell.

Basically, the speed of the browser depends upon the speed of the html parsing engine, available bandwidth, browser settings, speed of the cache and Javascript, just to mention the main variables.

Still, I'm interested how comes Opera's Javascript is so fast compared to the other browsers.

Re:Article is very misleading - JS benchmark only

"Still, I'm interested how comes Opera's Javascript is so fast compared to the other browsers."

Well, they didn't test it against WebKit/Safari/Konq, which blazes through Javascript tests. Firefox's Javascript engine (SpiderMonkey) leaves a lot to be desired, and well, Internet Exploder is just plain terrible at everything. Things will get better for Firefox once Mozilla figures out a way to integrate Tamarin, but this is still a while off.

Re:Article is very misleading - JS benchmark only

Well, they didn't test it against WebKit/Safari/Konq, which blazes through Javascript tests

It may blaze through tests, but in real life Konq is considerably slower than Firefox. I have to deal with a number of javascript ladden juggernauts like the ex-PeopleSoft eBusieness suite on a daily basis and konq is visibly much slower than Firefox.

The WebKit implementation is superior IMO

Testing javascript with Safari is like testing javascript with NoScript -- of course it's going to be faster since it doesn't really work

The WebKit implementation of JavaScript is easily better than the implementation in FireFox, Opera or Internet Explorer's JScript. Note I'm not saying just 'different', actually better - by which I mean it's demonstrably faster, is more feature complete, and requires less workarounds when you start doing complicated things (all centered around event handling though really, both FireFox and IE have issues with what you can/can't do when it comes to events and referencing properties of objects - in Safari everything I would expect to work, just does, though YMMV).

I've written both simple demos [iaincollins.com] and fairly sophisticated JavaScript apps [google.co.uk] (which can do Sim City / Civilization 2.5 isometric views like this [googlegroups.com] - and render them extremely quickly so you that you can pan around the environment as if it was a native title)).

When it comes to looping through a large array of arrays (e.g. the terrain tile detail in one of the above examples), applying style or class attributes to DOM elements, creating or moving DOM elements on a page and dealing with event handlers Safari wins hands down, followed by FireFox, Opera and IE (in all respects). The "Opera is the fastest" claim holds very little weight with me having compared them. What Opera has is a very fast UI that's extremely responsive, which is all a bit smoke and mirrors really. It's not particularly fast at script execution or object manipulation as soon as things get interesting (it lags behind Safari and FireFox certainly, but it's still far ahead of IE), and of course it renders perfectly valid pages very differently from Safari and FireFox (for which is sometimes possible to blame ambiguities in the standards, but that it doesn't follow the lead of Gecko/KHTML/Webkit or IE is a bit annoying - though do I appreciate the complexity involved).

Re:Article is very misleading - JS benchmark only

Well it depends at what, i had an MD5 routine and benchmark in javascript that was laughably slow on konqueror/safari...
The benchmark is at:
http://pentestmonkey.net/jsbm/index.html [pentestmonkey.net]

And i get the following results on a macbook pro 2.16ghz core2 duo running osx 10.4.10:

Safari (2.0.4):
MD5 Benchmark took 15.136 seconds for 3000 hashes (198 hashes/second)
MD4 Benchmark took 10.876 seconds for 2700 hashes (248 hashes/second)
SHA1 Benchmark took 19.052 seconds for 1900 hashes (100 hashes/second)

Camino (1.5.1):
MD5 Benchmark took 1.78 seconds for 3000 hashes (1685 hashes/second)
MD4 Benchmark took 1.271 seconds for 2700 hashes (2124 hashes/second)
SHA1 Benchmark took 1.931 seconds for 1900 hashes (984 hashes/second)

Firefox (latest nightly build):
MD5 Benchmark took 1.867 seconds for 3000 hashes (1607 hashes/second)
MD4 Benchmark took 1.299 seconds for 2700 hashes (2079 hashes/second)
SHA1 Benchmark took 2.077 seconds for 1900 hashes (915 hashes/second)

Firefox (2.0.5):
MD5 Benchmark took 2.628 seconds for 3000 hashes (1142 hashes/second)
MD4 Benchmark took 1.919 seconds for 2700 hashes (1407 hashes/second)
SHA1 Benchmark took 2.872 seconds for 1900 hashes (662 hashes/second)

Opera 9.23 (current stable):
MD5 Benchmark took 4.561 seconds for 3000 hashes (658 hashes/second)
MD4 Benchmark took 3.163 seconds for 2700 hashes (854 hashes/second)
SHA1 Benchmark took 4.812 seconds for 1900 hashes (395 hashes/second)

Opera 9.50 alpha (build 4404):
MD5 Benchmark took 1.446 seconds for 3000 hashes (2075 hashes/second)
MD4 Benchmark took 1.021 seconds for 2700 hashes (2644 hashes/second)
SHA1 Benchmark took 1.607 seconds for 1900 hashes (1182 hashes/second)

Quite impressive the improvements that have been made in the latest opera... Also, camino wasn't faster than the firefox nightlies last time i tried it (camino 1.0.4)...
I don't have access to msie or konqueror, i would assume konqueror performance would be similar to safari tho.

Re:Article is very misleading - JS benchmark only

can we assume the lack of IE7 benchmarks are due to being unfinished yet?

Re:Article is very misleading - JS benchmark only

Here's some results on Mac OSX (MacBook Pro Core Duo 2GHz):

Prerelease builds:

Safari 3 Nightly 177ms
Opera 9.5 Alpha 278ms
Firefox 3 Nightly 823ms

Production builds:

Safari 2 423ms
Opera 9.2 684ms
Firefox 2 880ms

Looks like Safari wins this one.

Re:Article is very misleading - JS benchmark only

I think the reason they're focusing on Javascript here is because that's a major optimization that took place in Opera 9.5. Actually, the changelog tells that they rewrote the ECMAScript engine. But Opera also had optimizations done to its table renderer, and due to the still all too frequent table layouts on the web, even used by modern web designers, it would be interesting to see more general tests of loading times etc. Opera would probably still come out very close on top though, as it has before in the pre-9.5 versions too.

Re:Article is very misleading - JS benchmark only

Not to mention font rendering. If using sub-pixel anti-aliasing, or anti-aliasing against the real background and not the document's bgcolor (or css equivalent), yes, it takes a lot longer, for a much better rendered result. Opera can be downright ugly when using small serif fonts on a non-uniform background, and Safari tends to dither against the wrong colour, especially if a table cell has a different colour than the document itself.

Regarding text rendering... What bugs me is that since the first Firefox, every so often, you get a horisontal line which is skewed by one pixel. This happens on both Linux and Windows, on different machines, with different fonts, with all Gecko engines. When this happens between lines, it's not TOO bad -- it just looks odd when there's suddenly a pixel more space between two lines than all the others, but when it happens in the text itself, it's VERY noticable. And if you select the text on that line and unselect it again, the problem goes away. It's like the rendering engine pre-calculates how much vertical space to set aside for the text in order to to increase rendering speed. Then, when drawing the text, the actual result never matches the space, so it duplicates or chops lines at random intervals until it the text fits. I'd rather wait a little longer and avoid this problem.

Opera faster _with JavaScript_

Okay, so Opera is probably a bit faster than Firefox in page rendering as well if they're faster at JavaScript, but the actual quote (emphasis mine) is:

When running various JavaScript speed tests, Opera 9.5 scored slightly higher (281ms) than the previous released version, 9.23 (546ms)

So Opera is much faster than FF when running JavaScript tests, according to Ars Technica.

Numbers are meaningless without context ;)

I'm fed up with the anti-Opera crap here...

Opera is faster than Firefox across the board. Always has been, and probably always will be. Put that into context whatever way you want. So what's the point of your emphasis again?

At the same time, Opera is also smaller, lighter, more stable, more innovative, better integrated, and comes from a company that behaves ethically towards the rest of the software community (eg, it does not engage in patent warfare to pummel the competition).

Yet because it's not open source (it's been "free as in beer" for quite some time now, but even that's news to some people here) it's practically awarded pariah status by many Firefox zealots who typically use nothing more than ignorance and FUD to put it down.

Seriously, the amount of anti-Opera, pro-Firefox propaganda (for want of a better word) here on Slashdot is ridiculous. Opera is, and always has been, a top-notch product.

In the eyes of this humble observer, it's a far better browser than any other, but regardless of our personal preferences, isn't it time that people gave it due respect? Or is good software engineering only to be appreciated if it comes from the open source community?

Re:I'm fed up with the anti-Opera crap here...

If security is what concerns you then, unless I'm very much mistaken, Firefox has had more vulnerabilities than Opera.

The whole "rigorously security audited" argument is a fallacy, unless you truly believe that Opera is somehow doing something that it shouldn't be doing. And that fallacy is blown out of the water when you realise that there isn't a single demonstrable example of Opera doing something as unethical as "phoning home" with your browsing habits, etc.

Look around. The minute that something like that happens, whether it's Microsoft, Real, Sony or whoever, it's exposed almost immediately. Why, then, do people maintain this "ooh, they could be doing something naughty" line about Opera, when the company has gone out of its way to be a positive member of the software community? It's FUD, pure and simple.

Look elsewhere on this story. You have people claiming that it's not "free as in beer". That's ignorance. You have people claiming it's not as fast as Firefox. That's ignorance again. You have people claiming that it might
be useful if only it would perform well on machines that are only equipped with 256MB. That's... well, do you want to guess what that is? Go on, guess. You have people bleating "big deal, speed doesn't matter". Yet these are the same people who bleat about how Firefox is better than MSIE because it's faster and less bloated.

It's all FUD and ignorance, FUD and ignorace. What happened to fair judgement and common sense?

Opera is a great product from a great company. Pure and simple.

Re:I'm fed up with the anti-Opera crap here...

I tried Opera.

Good browser it may be, but I don't like it. It's better than IE, but then, so is Lynx.

I like Firefox not so much for its speed (I'll admit Opera is faster), but for the extensions.

And yes, some of the more often used extensions do come off as copies of stuff first introduced in Opera, which makes Opera a bit of the Apple of the browser world.

And JFTR: Opera fanboys (the few that I've encountered) are worse than Linux, Mac and Amiga fanboys combined.

Re:I'm fed up with the anti-Opera crap here...

I would also like to say 2 words. "Web Developer". That's all. As a web developer, the web developer plugin makes web development so much easier. If there's a rendering bug, or something else on Firefox, then I don't worry about it too much, because I know it will be easy to fix. Change a cookie value, see hidden form values, edit HTML and CSS and see the results instantly, without reloading the page. I know that there's "web developer" plugins for IE and such, but I have yet to see one with the functionality and ease of use of the firefox one. And that is the reason I'll continue to use Firefox as my main browser, until something beats them on this front.

Re:I'm fed up with the anti-Opera crap here...

I used Firefox for about 4 years, and installed Opera this summer to do some testing on it.

Since then, I've used Opera for browsing and Firefox for web development. There's just no comparison between the two. And now that one of the other responses to this post has pointed me at this [opera.com], I may not use Firefox for anything other than testing in Firefox.

Of course, I'm one of those sufferers of the Firefox bug that causes it to use ridiculous amounts of memory. I've got a Firefox window open with Gmail (alas, Gmail breaks in Opera for me when composing mail), and it's consuming 180MB. I've got 2 opera windows open with about 15 tabs in one, including a few large Slashdot discussions, and it's consuming 120MB. So for me, there was no question when choosing between the two for everyday use.

--Jeremy

Re:Opera faster _with JavaScript_

You've clearly never used Opera if you're attempting to spin this article by claiming that we just plain don't know that Opera renders stuff in general near the top of the pack already, and also is perhaps the most standards compliant browser.

Not to mention that Opera 9.x is one of the only stable browsers with tentative support for HTML 5.

I get a kick out of FF fans on this site. FF is by no means bad, but Opera clearly has areas where it consistently outshines the open-source browser. Before, people used to say "I don't like ads in my browser" as an excuse for not using it. Then when it became free, it was "I use lots of GreaseMonkey scripts", despite the fact that you can use most GM scripts in Opera too.

Opera leads the way for most browsing achievements, and they show no signs of stopping. I've been using it since version 6, and though I give FF a whirl every .x build, I still have yet to see anything on FF that makes me believe it's worth the switch... and to top it off I'm a web developer by trade. I code for Opera, then break it for FF and IE.

Those numbers mean nothing...

without units. 281ms per what? Apparently a bunch of tests listed on http://celtickane.com/projects/jsspeed.php [celtickane.com]

Now my question is, how significant is ~500 ms for these tests? All I care about is how long it takes to load a typical webpage I surf, and for me, Firefox seems almost instantaneous for most pages. "Smacks silly" my be an overstatement.

Resource-conservation, not speed

Right now, the biggest issues with both IE and Firefox is a huge memory footprint. If Opera wants to bring something valuable to the table, make sure it can run smoothly on XP with 256 megs of memory. That would be valuable for a lot of people with aging hardware.